J Bus Ethics
DOI 10.1007/s10551-016-3225-4
The Impact of Corporate Social Responsibility Disclosure
on Corporate Reputation: A Non-professional Stakeholder
Perspective
Anastasia Axjonow1 • Jürgen Ernstberger2 • Christiane Pott1
Received: 27 February 2016 / Accepted: 22 May 2016
 Springer Science+Business Media Dordrecht 2016
Abstract This paper examines the impact of corporate
social responsibility (CSR) disclosure on corporate repu-
tation as perceived by non-professional stakeholders. Pro-
ponents of CSR disclosure argue that CSR disclosure can
be considered as a tool for reputation management. We
empirically investigate this claim using a reputation index
which tracks the general public’s perceptions of corporate
reputation over time. In our analysis, we focus on disclo-
sure in stand-alone CSR reports and control for CSR per-
formance. We find that, in contrast to the common belief,
stand-alone CSR reports do not influence corporate repu-
tation among non-professional stakeholders. However, we
are able to document that stand-alone CSR reports influ-
ence corporate reputation among professional stakeholders.
We also provide some evidence that transparent CSR dis-
closure on corporate websites can influence corporate
reputation among non-professional stakeholders.
Keywords Corporate reputation  Corporate social
responsibility disclosure  Non-professional stakeholders’
Data Availability Data are available from commercial databases and
public sources identified in the paper.
& Christiane Pott
Christiane.Pott@tu-dortmund.de
Anastasia Axjonow
Anastasia.Axjonow@tu-dortmund.de
Jürgen Ernstberger
ernstberger@tum.de
1 governance performance including preservation of environment,
TU Dortmund University, Friedrich-Wöhler-Weg 6,
44227 Dortmund, Germany
2
Technical University of Munich, Arcisstr. 21, 80333 Munich,
Germany
perceptions  Professional stakeholders  Voluntary
disclosure
JEL Classification L14  M14  M41
Introduction
A major goal of firms investing in corporate social
responsibility is to improve their reputation (KPMG
International 2011). Most firms now disclose information
on their CSR performance to make it more visible to
stakeholders (KPMG International 2011, 2013).1 Hence,
CSR disclosure can be considered as a tool for reputation
management and enhancement (Bebbington et al. 2008;
GRI 2016). While some prior studies document such a
notion for professional stakeholders (Toms 2002; Cho et al.
2012; Pérez et al. 2015), surprisingly little is known about
how disclosing CSR information influences corporate
reputation among non-professional stakeholders. This
paper brings evidence to this issue in an attempt to fill this
gap in prior literature. We investigate whether corporate
reputation among non-professional stakeholders like (po-
tential) consumers, employees, and the general public
changes around firms’ publication of stand-alone CSR
reports.
Reputation can be defined as ‘‘a perceptual representa-
tion of a firm’s past actions and results that describes the
firm’s ability to deliver valued outcomes to multiple
1
In general, CSR disclosure covers firms’ environmental, social, and
improvement in labor welfare, protection of human rights, contribu-
tion to the society, and pursuit of product safety providing non-
financial information besides financial information in annual reports
(Dhaliwal et al. 2011).
123

stakeholders. It gauges a firm’s relative standing both
internally with employees and externally with its stake-
holders, in both its competitive and institutional environ-
ments’’ (Fombrun and van Riel 1997, p. 10). For firms,
great importance is attached to reputation, since a good
reputation may have many favorable consequences such as
increasing profitability, enabling firms to charge premium
prices, attracting consumers and investors, or enhancing
access to capital markets (Fombrun and Shanley 1990;
Roberts and Dowling 2002; Rindova et al. 2005; Barone
et al. 2014). Hence, reputation is associated with numerous
potential benefits and is often seen as one of the firms’ most
important intangible assets, making it especially crucial for
companies to maintain or develop a good reputation
(Fombrun 2001; Barone et al. 2014).
Since CSR reports include information about a firm’s
CSR activities and, thus, provide information about doing
good, this could lead to stakeholders being able to better
identify themselves with the company, what could create a
greater commitment to the firm leading to an improved
reputation (Sen and Bhattacharya 2001; Bebbington et al.
2008). In contrast, critics doubt these benefits and consider
CSR disclosure as a waste of resources (Izzo and Mag-
nanelli 2012; Moser and Martin 2012).
In our analyses, we proxy for corporate reputation
among non-professional stakeholders using the BrandIndex
provided by the YouGov Group.2 This index measures the
corporate brand image and is based on a daily survey of
5000 randomly selected persons from the U.S. general
population. We use different methodological approaches to
investigate the effects of CSR disclosure on reputation
among non-professional stakeholders. First, we examine
the long-term effects by regressing our daily reputation
measure on CSR disclosure and several control variables
including CSR performance. Following Dhaliwal et al.
(2011, 2012, 2014), we measure CSR disclosure by a
dummy variable indicating whether a firm issues a stand-
alone CSR report. Second, we examine the short-term
effects by applying an event study type approach. We
analyze changes in the reputation index in different event
windows (3/2/1 months and 10/5 days) around the publi-
cation date of a CSR report. To control for a potential self-
selection bias and time-trends, we match the treatment
group of CSR reporters with a benchmark group of non-
CSR reporters.
Moreover, according to Du et al. (2010), stakeholder
groups differ from each other concerning their information
needs and the information channels they use making it
essentially important to differentiate between professional
2
The BrandIndex was already used by Luo et al. (2013) for
measuring brand dispersion and by Luo et al. (2014) and Panico et al.
(2014) for measuring reputation.
123
A. Axjonow et al.
and non-professional stakeholders and the CSR informa-
tion sources they use. For example, non-professional
stakeholders may be more informed about a company’s
reporting activities and also more interested in additional
information explicitly provided by CSR reports than con-
sumers or the general public. Non-professional stakehold-
ers might simply not know that a firm publishes a CSR
report or not be able to deal with the overwhelming amount
of information provided in such reports (Parguel et al.
2011). Thus, CSR information provided by another source
could be more appropriate when addressing non-profes-
sional stakeholders. Hence in further analyses, we also
consider reputational assessments of professional stake-
holders, measured by the Fortune magazine’s annual
America’s Most Admired Companies ranking3 and the
firm’s websites as an alternative CSR information channel,
since stakeholder group and CSR disclosure channel could
play an important role in influencing reputational
perceptions.
Our study contributes to the literature in several ways.
First, we contribute to the CSR reporting literature ana-
lyzing the impact of stand-alone CSR reports on reputation
among non-professional stakeholders, since prior studies
focus on financial effects of CSR reporting or reactions of
capital markets (e.g., Clarkson et al. 2013; Dhaliwal et al.
2011, 2012, 2014; Plumlee et al. 2015). Hence, prior
research on the effects of stand-alone CSR reports on
corporate reputation is sparse and does not include other
stakeholder groups than professional stakeholders.
Second, we contribute to the literature on the determi-
nants of corporate reputation. Some empirical studies
which sought to understand the determinants that influence
reputation provide evidence that, for instance, financial
factors like profitability, firm characteristics like size
(Fombrun and Shanley 1990), market actions (Basdeo et al.
2006), and special business activities like CSR activities
have an impact on reputation (Fombrun and Shanley 1990;
Williams and Barrett 2000).
Third, our study might be of interest to standard setters
and practitioners. One of the most often mentioned reasons
for CSR disclosure stated by reporting companies is that it
generally leads to a reputation improvement (KPMG
International 2011). However, mostly, these assertions are
accompanied by theoretical assumptions taking a positive
relationship between CSR disclosure and reputation for
granted, whereby empirical evidence concerning these
claims is missing. Our investigation provides a unique
empirical analysis of non-professional stakeholder per-
ceptions, whereas the YouGov BrandIndex allows us to
examine daily firm assessments. Hence, our study
3
This ranking was already used in several studies as a measure for
reputation (e.g., Fombrun and Shanley 1990; Basdeo et al. 2006).
The Impact of Corporate Social Responsibility Disclosure on Corporate Reputation…
represents an important empirical contribution to the the-
oretical discussion of the presented question. Moreover,
our investigation provides important practical implications.
Our findings indicate that it is important to differentiate
between professional and non-professional stakeholders
and the CSR information sources through which they
should be addressed. Hence, recommendations for firms
can be derived from our results. Especially, firms can draw
on the presented results when deciding where to publish
CSR information, regarding which stakeholder group they
want to address.
The remainder of our paper is organized as follows.
‘‘Related Literature and Research Question’’ section dis-
cusses the related literature and our research question.
‘‘Research Design’’ section presents a description of our
main variables, the sample selection process, and
methodology. ‘‘Results’’ section discusses the descriptive
statistics and the results of our empirical tests. ‘‘Additional
Analyses and Robustness Checks’’ section presents addi-
tional analyses and robustness checks. Finally, the last two
sections provide a discussion of the results, show limita-
tions, and conclude.
Related Literature and Research Question
There is a conspicuous increase in firms’ CSR disclosure.
The percentage of the largest 250 companies in the world
issuing stand-alone CSR reports grew from 35 % in 1999 to
almost 93 % by 2013 (KPMG International 2011, 2013). A
growing body of research investigates the determinants and
effects of CSR disclosure. More precisely, prior research
on CSR disclosure focuses on financial effects and reac-
tions of financial market participants (Dhaliwal et al.
2011, 2012, 2014; Clarkson et al. 2013; Griffin and Sun
2013; Matsumura et al. 2014; Cheng et al. 2015; Plumlee
et al. 2015). However, despite theoretical assumptions and
anecdotal evidence that CSR disclosure should enhance a
firm’s reputation, empirical evidence is scarce.
One area of research examines the impact of CSR
activities and performance on assessments of corporate
reputation providing evidence of a positive relation
(Fombrun and Shanley 1990; Williams and Barrett 2000;
Lev et al. 2010; Hur et al. 2014). Prior studies also show
that CSR practices positively influence consumers’ atti-
tude towards the company, thus, affecting its reputation
(Sen and Bhattacharya 2001; Berens et al. 2005; Luo and
Bhattacharya 2006). Furthermore, Linthicum et al. (2010)
examine the influence of CSR activities on firms’ repu-
tation among investors during a period of crisis (the
Enron audit failure) but do not find evidence that CSR
activities can burnish a firm’s reputation in a time of
depression.
It seems surprisingly that only few prior studies inves-
tigate the impact of environmental disclosures on assess-
ments of corporate reputation, since one of the reasons
mentioned most often by reporting companies for CSR
disclosure is that they associate a reputation improvement
with their reporting activities (cited by 67 % of the G250).
The GRI4 (2016) also claims that CSR reporting can lead to
brand and reputation enhancement.
In line with these assumptions, Toms (2002) provides
evidence for a sample of UK firms that environmental
disclosure quality in annual reports contributes signifi-
cantly to the creation of environmental reputation. This
study uses the corporate reputation rankings for the com-
munity and environmental responsibility aspect of Man-
agement Today’s survey of Britain’s most admired
companies in 1996 and 1997 providing the views of pro-
fessional stakeholder groups.5 Using a sample of 92 U.S.
firms from environmentally sensitive industries, Cho et al.
(2012) find that environmental performance is negatively
related to environmental reputation, whereas environmen-
tal disclosure extent (analyzed in annual reports and stand-
alone CSR reports) is positively related to environmental
reputation. Environmental reputation is measured by using
the ranking of the ‘‘greenest’’ companies in America for
2009 reported in the Newsweek magazine which incorpo-
rates the views of environmental experts. In a more recent
study, Pérez et al. (2015) provide evidence that for the
financial industry CSR reporting is more closely linked to
their CSR reputation (among expert groups) than the CSR
reporting of basic, consumer goods, and services industries.
They analyze Spanish companies and use the ‘MercoEm-
presas Responsables’ index as a measure for CSR
reputation.
In contrast to prior literature, this paper focuses on the
impact of CSR disclosure on corporate reputation based on
the perceptions of non-professional stakeholder groups.
Information provided via CSR disclosure does not only
address professional stakeholder groups or experts but also
other stakeholder groups such as consumers and employ-
ees. These groups are important addressees because they do
not seek CSR information in annual reports because of its
technical nature (de Villiers and van Staden 2011). More-
over, we investigate a broad concept of reputation among
the general public, instead of focusing on just a certain part
of reputation among specific experts. Fombrun and Shanley
(1990) criticize that considering one dimension of
4
The Global Reporting Initiative (GRI) is an organization in the
sustainability field. GRI promotes the use of sustainability reporting.
It has developed a sustainability reporting framework including
reporting guidelines to enhance organizational transparency and
accountability.
5
Hasseldine et al. (2005) complement and extend the work of Toms
(2002) using more recent data and find corroborating results.
123

reputation leads to disregarding the correlations among
other reputation dimensions. Since the dimensions are
conceptually distinct and show considerable relatedness,
considering only one dimension (e.g., the social responsi-
bility dimension) may not be sufficient. Moreover, we
examine a broader concept than environmental disclosure,
namely CSR disclosure. Thereby, we respond to Fombrun
and Shanley’s (1990), Moser and Martin’s (2012), and
Huang and Watson’s (2015) recent calls to investigate the
impact of CSR disclosure on other audiences firms interact
with, not just financial stakeholders.
From a theoretical perspective, CSR disclosure can be
seen as a signaling tool that provides additional informa-
tion to stakeholders for attending to information asymme-
try (Akerlof 1970; Watts and Zimmerman 1986). Firms
that make considerable efforts in the CSR area have an
incentive to provide voluntary information about these
attempts. This implies that firms which have a good CSR
performance want to signal that fact to stakeholders by
voluntarily issuing reports and, thus, distinguishing them-
selves from bad CSR performers (Akerlof 1970). However,
firms with a poor performance have the incentive not to
issue a report or make less disclosure (Verrecchia 1983;
Dye 1985; Healy and Palepu 2001). Empirical research
provides evidence that is consistent with economic theory
delivering results that firms with better (worse) CSR per-
formance are more (less) likely to provide CSR disclosure
(Clarkson et al. 2008; Eccles et al. 2014). In our reputation
context this means that CSR disclosure could be considered
by the public as a positive signal that a firm makes a par-
ticular effort in CSR activities and demonstrates a good
CSR performance. This again might influence the public’s
perception of a firm and, thus, affect its reputation posi-
tively. Sen and Bhattacharya (2001) claim that investing in
CSR activities leads to the point that stakeholders can
better identify themselves with the company. This creates a
greater commitment to the firm leading to better evalua-
tions of the organization. Overall, both theoretical and
empirical evidence supports a significant association
between CSR activities and reputation. This association, in
turn, suggests issuing non-financial disclosure, such as
those concerning CSR activities, could also have an impact
on reputation.
CSR disclosure can also be considered as a communi-
cation tool, since it answers numerous stakeholder
demands and firms use this medium to report in line with
their key stakeholders’ expectations (Sweeney and
Coughlan 2008). Using this communication instrument,
companies are able to represent themselves in a self-
laudatory way by reporting about their efforts in doing
good (Hooghiemstra 2000). Such self-presentation via CSR
disclosure allows firms to construct a certain social identity
in order to be regarded as socially responsible. This
123
A. Axjonow et al.
identity construction process described by Scott and Lane
(2000) involves managers of an organization impressing
desired images on their stakeholders. Thus, CSR disclosure
can be understood as an impression management and
identity construction strategy. Impression management
belongs to the field of social psychology and is defined by
Schlenker (1980) as ‘‘the conscious or unconscious attempt
to control images that are projected in real or imagined
interactions.’’ Hence, by interacting with stakeholders via
CSR disclosure, firms can convey certain desirable and
favorable impressions. This could influence people’s per-
ceptions of the company in a positive way, thus leading to
an enhanced reputation.
Besides arguments why CSR disclosure could influence
corporate reputation positively, there are also some
opposing opinions. Critics of CSR disclosure claim that
even if such disclosure responds to stakeholders’ demands,
this is potentially achieved at the shareholders’ expense
(Moser and Martin 2012; Martin and Moser 2016). These
critics argue that investment in CSR can be seen as a waste
of resources (Izzo and Magnanelli 2012). In our reputation
context, this could mean that if on average stakeholders
have this negative association with CSR disclosure and see
investment in such a kind of disclosure as a waste of
resources, CSR disclosure would not be seen as a positive
signal. Instead, stakeholders could get the impression of a
poor distribution of financial resources. This circumstance
could lead to no reputation improvement or even reputation
erosion.
Another explanation of CSR disclosure having no
effects on corporate reputation is the existing uncertainty
about the credibility of such disclosure. Since CSR
reporting is voluntary, the reports are not required to be
verified by auditors (Simnett et al. 2009). Hence, managers
are able to give the information they provide a more pos-
itive slant (Moser and Martin 2012). Thus, stakeholders
could have doubts concerning the reliability of CSR dis-
closure with the result that they would not be easily
influenced by the issuance of a CSR report and the infor-
mation provided there. This again leads to the result that
the issuance of CSR reports would have no impact on
reputation.
Socio-political theory suggests that social and political
pressure lead managers of poorly performing firms to dis-
close more to legitimize their organization’s actions (Patten
2002; Cho and Patten 2007). If stakeholders get the
impression that CSR disclosure is issued to lessen public
pressure by fulfilling their expectations rather than to
provide information about an actually good CSR perfor-
mance, one might expect that the issuance of a CSR report
would have no impact on reputation.
Godfrey (2005) argues that a firm has to meet two cri-
teria for a reputational act, which in our context is the
The Impact of Corporate Social Responsibility Disclosure on Corporate Reputation…
issuance of a stand-alone CSR report, to generate positive
reputational capital. First, there has to be consistency
between the underlying ethical value of the act and the
ethical values of the community. Second, the act should not
be perceived as an ingratiating attempt to generate public
favor. Godfrey (2005) claims that only acts that are asso-
ciated with the firm’s underlying strategy can enhance
reputation, whereas actions judged as ingratiating attempts
could even lead to a deterioration of reputation. Hence, if
CSR disclosure is perceived by stakeholders as a tool for
self-promotion to gain public favor and as acting in a self-
serving manner for putting the firm in a better position, it
could have no impact on the attractiveness of an organi-
zation. This again would lead to no (positive) effect on a
firm’s reputation.
Finally, non-professional stakeholders could simply not
know that a firm publishes a CSR report or be over-
whelmed with the amount of information provided in such
reports (Parguel et al. 2011). This stakeholder group can
also demonstrate other information needs than professional
stakeholders like financial analysts or institutional inves-
tors and, thus, respond differently to CSR reporting (Du
et al. 2010). While firms might tend to cater to professional
stakeholder groups, the issuance of CSR reports might not
have any impact on non-professional groups and no
improvement of their assessments of corporate reputation
would arise.
To conclude, the issuance of a stand-alone CSR report
does not necessarily have an effect on a firm’s reputation.
Given the different predictions of various theories, the
impact of CSR reporting on corporate reputation among
non-professional stakeholders is ultimately an empirical
question. We state following research question:
Research Question: Does the issuance of a stand-alone
CSR report have an effect on a firm’s reputation among
non-professional stakeholders?
Research Design
Main Variables
Measure for Corporate Reputation
Following Sarstedt et al. (2013) who claim that future
research should consider different methods and measures
for reputation, we use the BrandIndex provided by YouGov
Group for our corporate reputation measure.6 The YouGov
6
For a listing as well as a more extensive description and analysis of
further reputation measures, see Schwaiger (2004), Bebbington et al.
(2008), and Sarstedt et al. (2013). The BrandIndex was already used
by Luo et al. (2013) for measuring brand dispersion and by Luo et al.
(2014) and Panico et al. (2014) for measuring reputation.
Group specializes in online-panels monitoring brands in
the United States. According to marketing literature, most
definitions of corporate reputation are based on companies’
brand attributes (Fombrun and van Riel 1997; Rindova
et al. 2005; Nikolaeva and Bicho 2011). Following Fom-
brun (2001), reputation is analogue to brand equity. Hence,
the brand evaluations can be seen as a proxy for the rep-
utation of the firms to which the brands belong. We match
the brands covered by YouGov with the brand owning
firms for which the company name equals the brand name
to ensure that respondents evaluating the brand are
informed about the brand owner (Larkin 2013). Thus,
evaluating firm name equal brands, the participants auto-
matically evaluate the company which is behind the brand.
YouGov monitors about 1100 brands in 20 different
industry sectors for the U.S. market surveying 5000 ran-
domly selected persons from the U.S. general population
(from a panel size of 1,800,000 persons) out of all relevant
demographic groups, on a daily basis. This large panel size
is beneficial because it can represent the firm’s stakehold-
ers in a better way. Since the respondents of the surveys
belong to the general population (comprising also (poten-
tial) employees and (potential) costumers), non-profes-
sional stakeholder groups can be considered. Since many of
the social and environmental issues contained in CSR
reports address aspects associated also with employees or
consumers, it is important to assess whether and how these
groups react to the issuance of CSR reports (Moser and
Martin 2012; Huang and Watson 2015). Another advantage
of the BrandIndex is that it is available on a daily basis.7
Thus, perceptions of the firm’s stakeholders and changes in
these perceptions can be reflected more timely (Luo et al.
2013).
The BrandIndex is an aggregate index of six dimensions
which are represented by six questions. The dimensions
and questions are provided in Appendix Table 8. The
answers to the questions are collected by YouGov in the
following way. For each brand, three responses to the
questions are possible: positive, negative, and neutral.
First, for a given industry sector, the respondents choose all
brands for which they agree to the positive question (e.g.,
‘generally positive’ impression concerning the question
‘‘Of which brands do you have a ‘generally positive’ or
‘generally negative’ impression’’). Then, they choose all
brands for which they agree to the negative question (e.g.,
‘generally negative’ impression). The rest of the brands are
classified as neutral. Since respondents rate the competing
brands within one sector concurrently, brand competition
7
Nevertheless, we run a regression with a yearly BrandIndex
Reputation measure after calculating averages of the daily measures.
This analysis contains 465 firm-year observations and provides the
same main results as the analyses on a daily basis.
123

effects are controlled for (Luo et al. 2013). Moreover, in
order to reduce common method bias from the same survey
respondent, the brand assessments for each dimension are
measured separately across respondents. This means that
any respondent is asked about his or her perception of only
one brand dimension for a specific industry but not all six
dimensions for the same sector, whereby the dimension-
industry combination is randomized.
For each of the six dimensions, the daily brand rating
scores are calculated. This is done by taking the differences
of the number of respondents who agree with the positive
question and the number of respondents who agree with the
negative question divided by the sum of positive and
negative respondents (omitting neutral respondents)
allowing the scores to vary between -100 and 100. Neutral
respondents are omitted due to the possibility that a high
number of neutral assessments could indicate that the brand
is not known to the respondent, rather than that he or she
evaluates the brand neutrally, which could potentially bias
the BrandIndex scores. An aggregation of the daily scores
of the six dimensions leads to the daily BrandIndex. As a
consequence, this index comprises daily evaluations taken
by the respondents and does not consist of just one eval-
uation at the end of the year, as it is the case with other
reputation measures.8
Hence, we generate a daily corporate reputation score
(REPUTATION) as proxy for reputation for each firm in
our sample.
Measure for Corporate Social Responsibility Disclosure
Following Dhaliwal et al. (2011, 2012, 2014), we use a
binary variable that indicates whether firms publish stand-
alone CSR reports. By issuing stand-alone CSR reports,
companies demonstrate a special attempt to make such
information available, whereby stakeholders construct
reputations from obtainable information about firms’
activities (Fombrun and Shanley 1990).
In line with Simnett et al. (2009), we obtain information
about which firms in our sample have issued stand-alone
CSR reports in the regarded period from CorporateRegis-
ter.com (http://www.corporateregister.com). Corpo-
rateRegister.com is the global corporate responsibility
resources website that hosts the world’s most comprehen-
sive directory of corporate non-financial reporting.
8
Using a measure based on a single evaluation at the end of the year
might raise concerns about endogeneity occurring from unknown
events prior to the evaluation date. Thus, we believe that our index is
more appropriate to reflect the public’s perceptions of a firm, avoiding
endogeneity.
123
A. Axjonow et al.
Sample
YouGov provides daily data for 1100 brands available for the
U.S. market surveyed between January 1, 2010 and
December 31, 2012 which represents our sample period. We
refrained to analyze assessments after 2012, since YouGov
changed the methodology for determining the BrandIndex
which could potentially bias our results. The sample com-
prises single-brand and multiple-brands firms. Our initial
sample consists of 1,204,500 brand-day observations for
which the BrandIndex is available. We exclude observations
of brands which could not be assigned to a firm because of an
acquisition or merger process during the observed period.
We also delete observations from brands that do not belong
to U.S. companies. Further, we delete brand-day observa-
tions of brands with names not equal to the company name to
ensure that respondents evaluating the brand are informed
about the brand owner (Larkin 2013) and can match the
brand with a stand-alone CSR report of the firm.9 Hence,
each firm is represented by the BrandIndex of one name-
equal brand and the number of brand-day observations is,
thus, identical to the number of firm-day observations. We
exclude 113,378 firm-day observations of firms not listed in a
regulated stock market. We also exclude all observations
with missing values for the dependent or explanatory vari-
ables (283,597 firm-day observations), whereby financial
data for the explanatory variables was obtained from
Thomson Reuters Datastream. Our final sample consists of
114,296 firm-day observations from 164 individual firms.
The sample selection process is summarized in Table 1.
Empirical Model
To address our research question, we use the following
OLS regression which regresses our reputation proxy on
CSR disclosure and several control variables, estimating
the following regression model:
REPUTATIONi;t ¼ b0 þ b1 CSR REPORTi;t
þ b2 CSR PERFORMANCEi;t1
þ b3 SIZEi;t1 þ b4 ROAi;t1
þ b5 MtBi;t1 þ b6 YIELDi;t1
þ b7 RISKi;t1 þ b8 INSTITUTIONALi;t1
þ b9 DIVERSi;t1 þ b10 ADVERTISINGi;t
þ b11 MEDIA EXPOSUREi;t þ b12 LEVi;t1
þ b13 LITIGATIONi;t1 þ b14 R&Di;t1
þ industry fixed effects þ ei;t :
ð1Þ
9
Main results remain constant when we do not delete brands with
names not equal to firm name running the main regressions with all
brands.
The Impact of Corporate Social Responsibility Disclosure on Corporate Reputation…
Table 1 Sample selection
Step and respective procedure
Selection on brand level
Starting sample: Brand-day observations for which YouGov provides information from 2010 to 2012
Deleting not allocable brands
Deleting foreign brands
Deleting brands with names not equal to firm name
Deleting brands belonging to privately held firms
Deleting missing dependent variable
Deleting missing control variables
Final sample
This table shows the sample selection process of the final sample used for all major analyses containing 114,296 firm-day observations. After
deleting brands with names not equal to firm name, each firm is represented by only one brand. Hence, brand-day observations are analogue to
firm-day observations in the final sample
Definitions of all variables are provided in Appendix
Table 9. The dependent variable is the daily BrandIndex
value of our sample firms. CSR disclosure, as our variable
of interest, is proxied by, CSR_REPORT, an indicator
variable which is equal to 1 if a firm publishes a stand-
alone CSR report on day t and on up to 90 days (3 months)
after the publication date, and 0 on all other days.10 As
alternative time windows we use 60 days (2 months),
30 days (1 month), 10 days, and 5 days after the publica-
tion date.
Based on prior literature, we use several control vari-
ables.11 Prior literature shows that CSR performance could
influence reputation (e.g., Brammer and Pavelin 2006).
Hence, we include CSR_PERFORMANCE into our model.
CSR_PERFORMANCE is defined as the average of
10
We use different event windows, since the effect of a CSR report
publication should last some days or months after the publication. We
use quite large windows (3 and 2 months), since we investigate the
assessments of non-professional stakeholders and it is likely that these
groups may respond later to new information than equity market
participants. We desist from using an event window that exceeds 3
months because of potential confounding events. Our findings are
unchanged if CSR_REPORT is equal to 1 the whole year when a firm
publishes a CSR report in this year, and 0 otherwise, not forming
different event windows. Alternatively, we set CSR_REPORT the
whole year equal to 1 if a firm publishes a CSR, H&R, or
Sustainability report or publishes a section on CSR, H&R, or
Sustainability in the annual report which leads to the same main
results. Also, using a propensity score matching by size and industry-
year for dealing with a potential self-selection problem does not
change the tenor of our main results. Moreover, we run the regression
just for the CSR reporters, omitting observations of non-reporters, for
all event windows. The main results remain unchanged as well.
11
Mainly, we use lagged control variables because daily assessment
values included in our reputation measure are obtained during the
whole year and, thus, are more likely to be based on previous year
financial data.
Brand-day observations
1,204,500
-65,700
-129,804
-497,725
-113,378
-181,301
-102,296
114,296
environmental, social, governance, and economic scores
provided by the Thomson Reuter’s ASSET 4 database.12
We rely upon the seminal study by Fombrun and Shanley
(1990) in identifying additional control variables. We
include measures of firm size, financial performance, risk,
institutional ownership, diversification, advertisement, and
media exposure. Firm size (SIZE) is measured as the natural
logarithm of total assets. It controls for the effect that large
firms are more visible than small firms. Thus, the general
public might remember small firms less readily than larger
ones (Fombrun and Shanley 1990; Brammer and Pavelin
2006). Financial performance is taken into account by using
three measures. The return on assets (ROA) and the ratio of
market to book value (MtB) capture the beneficial effects of
high performance because stakeholders may view finan-
cially successful firms more favorably. The dividend yield
(YIELD), as the ratio of dividends to share price, comprises
opposing views, since it could have a positive as well as a
negative effect on reputation (Fombrun and Shanley 1990).13
As stakeholders tend to be risk-averse, firms that offer less
risk may exhibit the better reputation. Following Fombrun
and Shanley (1990), we include a measure for market risk
(RISK) gauged by the firm’s beta coefficient. INSTITU-
TIONAL captures the firm’s stock concentration in institu-
tional hands. Because a high amount of institutional
shareholders may signal to other stakeholders that the firm’s
performance is closely monitored and deemed to be
12
Most recent studies use the ASSET 4 dataset as measure for CSR
performance (e.g., Cheng et al. 2014; Dhaliwal et al. 2014; Eccles
et al. 2014; Lys et al. 2015). A more detailed description of the
ASSET 4 database is provided by Cheng et al. (2014).
13
On the one hand, high dividend payouts can signal a higher
profitability of a company positively affecting public’s perceptions of
that firm. On the other hand, taking a long-run view, high payoffs
could signal that the firm lacks attractive investment opportunities to
ensure future cash flows leading to no reputation enhancement
(Fombrun and Shanley 1990).
123

satisfactory by these well-informed and experienced inves-
tors, a higher institutional ownership might be viewed by the
stakeholders more favorably (Fombrun and Shanley 1990;
Brammer and Pavelin 2006). In addition, Fombrun and
Shanley (1990) argue that a firm’s strategy according to its
diversification posture can affect the public’s assessments
and, thus, a firm’s reputation. Contrary to focused firms,
unrelated diversification could make a negative impression
because informational signals about individual divisions
may be ambiguous and, thus, more difficult to interpret.
Therefore, we include the variable DIVERS, as the number of
business segments in which the firm operates, to control for
its diversification posture (Luo and Bhattacharya 2006).
Furthermore, we control for advertising intensity. We con-
sider that advertising aims at presenting firms in a favorable
light, which might have an impact on reputation (Fombrun
and Shanley 1990). We use ADVERTISING which is an
indicator variable that is equal to 1 if the regarded firm was on
the ‘‘100 Leading U.S. Advertisers’’ list in 2011 monitored
by Advertising Age, and 0 otherwise, to control for adver-
tising intensity. The list contains U.S. firms with the highest
advertising expenditures.14 Moreover, we include the vari-
able MEDIA_EXPOSURE, which is based on daily infor-
mation provided by YouGov indicating whether the
respondents have heard something positive or negative in the
REPUTATIONi;t ¼ b0 þ b1 CSR REPORTi;t þ b2 POSTi;t þ b3 CSR REPORTi;t  POSTi;t
X ð2Þ
þ bk CONTROLS þ industry fixed effects þ ei;t
media about the sample brands, since the publics’ assess-
ments could be affected by opinions in the media (Fombrun
and Shanley 1990).
We additionally control for leverage (LEV) because a
higher degree of leverage could lead to lower reputational
assessments, as it represents a strain upon future returns
that may threaten the viability of the firm (Brammer and
Pavelin 2006). LEV is defined as the ratio of total debt to
total assets. We also control for litigation risk (LITIGA-
TION) because firms that face a higher level of litigation
risk are likely to be assessed negatively by the stakehold-
ers. LITIGATION is an indicator variable which is equal to
1 if the firm operates in a high-litigation industry (SIC
codes of 2833–2836, 3570–3577, 3600–3674, 5200–5961,
and 7370), and 0 otherwise (Francis et al. 1994; Dhaliwal
et al. 2011). Moreover, we include research and
14
We assume that firms with a high advertising intensity in 2011
should also have high expenditures in the previous and following
years. Thus, ADVERTISING remains constant over time.
123
A. Axjonow et al.
development expenditure as the ratio to total assets (R&D)
because a firm’s R&D intensity enhances innovation
activities and therefore may affect stakeholders’ evalua-
tions of the firm (Luo and Bhattacharya 2006).15 Thus, an
investment in achieving higher product quality could
influence reputation positively (Brammer and Pavelin
2006).
In all specifications of the model in Eq. (1), we control
for industry fixed effects to capture differences between
assessments of various industries.16 Furthermore, we
cluster the standard errors by firm to control for
heteroscedasticity and correlation among the residuals
(Gow et al. 2010).17
In an alternative specification, we conduct an event-
based analysis for investigating our research question.
We use a differences-in-differences research design to
investigate whether there are significant changes in rep-
utational assessments following the publication of CSR
reports.18 Here again, we examine different event win-
dows (3/2/1 months and 10/5 days) around the publica-
tion date of a CSR report. To control for a potential self-
selection bias, we use a control sample of matched
benchmark firms. We use a 1:1 propensity score
matching by size and industry-year.19 This leads to fol-
lowing regression model:
where CSR_REPORT is an indicator variable which is
equal to 1 for the whole (yearly) period if a firm pub-
lishes a CSR report, and 0 otherwise. We add to the
15
R&D is set to zero if data on R&D expenditure were missing. Our
major findings are unchanged for all five event windows when we
eliminate firm-day observations with missing R&D data, resulting in a
sample of 59,561 firm-day observations.
16
We do not use year fixed effects because of the small observation
period of 3 years. Using also year fixed effects does not change our
main result.
17
Clustering by year or no clustering at all does not change the tenor
of our main result.
18
Since the reputation of our sample firms is highly volatile over
time, this event study type approach addresses the question whether
CSR report publication causes the reputational changes. The course of
the average firm reputation is provided in Appendix Fig 1.
19
This procedure allows us to allocate publication dates of the
treatment group (CSR reporters) to the matched pairs of the control
group (non-reporters). Our main results remain constant when we
match by CSR performance for further distinguishing actual CSR
activities from the disclosure of these activities.
The Impact of Corporate Social Responsibility Disclosure on Corporate Reputation…
previous equation POST, an indicator variable which is
equal to 1 if an observation relates to a post-publication
date in the regarded event window and 0 if it relates to a
pre-publication date in the regarded event window for all
matched pairs (e.g., POST equals 1 for the publication
date and for 90 firm-day observations after publication,
and equals 0 for 90 firm-day observations before the
report was published). Observations that exceed the
event window are excluded. The interaction of
CSR_REPORT and POST effectively measures the dif-
ference in changes in corporate reputation for the firms
that publish a stand-alone CSR report (treatment group)
relative to non-CSR reporters (control group).
In all specifications of the model in Eq. (1), we control
for industry fixed effects. Furthermore, we cluster the
standard errors by firm (Gow et al. 2010).
Results
Descriptive Statistics
Table 2 provides descriptive statistics for all variables used
in our main model (Eq. (1)). To mitigate the influence of
outliers, all the continuous variables are winsorized at the
1st and 99th percentile. Panel A of Table 2 shows a mean
value of 39.5 for our reputation proxy (REPUTATION).20
Please recall that reputation scores can vary between -100
and 100. Further, 37 % of the observations in our sample
have issued stand-alone CSR reports during the observation
period. Moreover, the firms in our sample have mean total
assets of 19,326.74 million $ (SIZE = 16.777).
Panel B of Table 2 reports descriptive statistics by
industry. We use the 12 industry classification by Fama and
French (2015). The most heavily represented industry is
Wholesale, Retail, and Services (37,314 firm-day obs.),
followed by Finance (25,537 firm-day obs.), whereas
Chemicals and Allied Products comprises the smallest
number of firm-year observations (1494 firm-day obs.).
The highest mean value for REPUTATION is achieved by
the Chemicals and Allied Products industry (86.39)21 fol-
lowed by Manufacturing (69.543). The industry with the
lowest reputation score is Finance with a mean value of
-3.84. The highest proportion of firms issuing a stand-
alone CSR report belongs to the Healthcare industry, fol-
lowed by Consumer Non-Durables. Firms that belong to
the Consumer Durables industry show the lowest propor-
tion of CSR report issuance.
20
The minimum (maximum) score amounts to -62.6 (93.6).
21
It must be noted that this industry is represented by a very low
number of observations which may cause this result.
Table 3 provides correlations among the variables
included in Eq. (1). The correlation between CSR_RE-
PORT and REPUTATION is significantly positive, but low
(Pearson 0.04; Spearman 0.04). Consistent with economic
theory on voluntary disclosure, the correlation between
CSR_PERFORMANCE and CSR_REPORT is also signifi-
cantly positive (Pearson 0.55; Spearman 0.57), indicating
that firms with superior CSR performance are more likely
to issue CSR reports to distinguish themselves from bad
performers. Overall, we do not find evidence of excessively
high correlations between the control variables. Thus, our
setting does not raise multicollinearity concerns.22
Main Results
Table 4 provides the results of estimating Eq. (1) for dif-
ferent model specifications. The first columns of each time
window provide results for Eq. (1). The estimated coeffi-
cient at CSR_REPORT is not statistically significant for
each event window. Thus, we find no significant relation
between CSR_REPORT and REPUTATION.
Regarding our control variables, we find significance in
the predicted direction for CSR_PERFORMANCE, MED-
IA_EXPOSURE, and R&D. All other explanatory variables
are not statistically significant.23 The adjusted R2 is 66 %
for each specification and is, thus, even higher than the
value reported by Toms (2002).
The second columns of Table 4 provide the results when we
include an interaction term of CSR_REPORT and an indicator
variable (HIPERFORM) that equals 1 when the firm’s CSR
performance is greater than the median CSR performance of
the firm’s industry peers, and 0 otherwise. This specification
allows us to analyze whether there is an effect of CSR reporting
on corporate reputation when firms have a high CSR perfor-
mance. All main results remain unchanged and the interaction
term is also not statistically significant.
In Table 5, we report the multivariate differences-in-dif-
ferences regression results for Eq. (2) for all five event
22
Moreover, the use of variance inflation factors does not lead to
concerns about multicollinearity, since all factors are below 10.0 and
most factors are below 5.0.
23
Although the financial performance variables in our main analysis
are not statistically significant, indicating that our reputation measure
does not suffer from the financial halo effect that Brown and Perry
(1994) documented with regard to the Fortune magazine’s annual
Most Admired Companies ranking, we run a regression with Brown
and Perry’s (1994) financial performance variables to ensure this.
Following Brown and Perry (1994), these variables are market to
book ratio, average return on assets, sales, growth (percentage change
in sales), and risk (debt/equity). We use lagged variables, since,
contrary to the AMAC ranking, our reputation measure is obtained
during the whole year. Since the adjusted R2 is very low, providing no
evidence for the existence of a financial halo effect, we use the
reported reputation measure in our analyses.
123
 A. Axjonow et al.

Table 2 Descriptive statistics

Panel A: Full sample (n = 114,296)
Variable Mean Std. 25 % Median 75 %

REPUTATION 39.477 38.401 14.200 48.200 70.700

CSR_REPORT 0.369 0.482 0.000 0.000 1.000
CSR_PERFORMANCE 63.619 28.288 36.990 71.440 91.340
SIZE 16.777 1.827 15.343 16.530 18.094
ROA 0.061 0.076 0.011 0.055 0.103
MtB 2.352 6.732 1.150 2.215 3.668
YIELD 0.513 0.650 0.000 0.250 0.740
RISK 1.149 0.611 0.790 1.070 1.350
INSTITUTIONAL 13.278 18.862 0.490 4.020 19.730
DIVERS 4.767 2.061 3.000 5.000 6.000
ADVERTISING 0.279 0.448 0.000 0.000 1.000
MEDIA_EXPOSURE 10.678 12.649 1.600 9.100 19.100
LEV 0.204 0.183 0.063 0.163 0.286
LITIGATION 0.383 0.486 0.000 0.000 1.000
R&D 0.013 0.031 0.000 0.000 0.003
Panel B: Descriptive statistics by industry for the main variables
Industry Statistics REPUTATION CSR_REPORT CSR_PERFORMANCE

Consumer non-durables N 5612 5612 5612

Mean 47.966 0.718 83.952
Median 65.000 1.000 95.060
Consumer durables N 2727 2727 2727
Mean 57.620 0.095 74.894
Median 61.300 0.000 92.580
Manufacturing N 6023 6023 6023
Mean 69.543 0.460 74.258
Median 72.000 0.000 77.930
Oil, gas, and coal extraction and products N 2925 2925 2925
Mean 14.244 1.000 89.723
Median 14.200 1.000 93.160
Chemicals and allied products N 1494 1494 1494
Mean 86.390 0.651 93.368
Median 87.100 1.000 96.170
Business equipment N 10,197 10,197 10,197
Mean 62.813 0.440 77.021
Median 72.400 0.000 88.900
Telephone and television transmission N 10,057 10,057 10,057
Mean 20.500 0.326 51.915
Median 19.000 0.000 46.610
Wholesale, retail, and some services N 37,314 37,314 37,314
Mean 53.915 0.260 55.813
Median 58.500 0.000 52.400
Healthcare, medical equipment, and drugs N 1981 1981 1981
Mean 62.990 0.754 94.490
Median 63.700 1.000 94.680
Finance N 25,537 25,537 25,537
Mean -3.840 0.391 63.515

123
The Impact of Corporate Social Responsibility Disclosure on Corporate Reputation…
Table 2 continued
Panel B: Descriptive statistics by industry for the main variables
Industry Statistics
Median
Other N 10,429
Mean
Median
This table provides descriptive statistics of the full sample for all variables used in our main model (Panel A) and descriptive statistics by industry
for the main variables (Panel B). We use industry classification by Fama and French (2015). Variables are defined as described in Appendix
Table 9. We winsorized all continuous variables at the 1st and 99th percentile. Panel A and Panel B include 114,296 firm-day observations
windows.24 For each event window, the coefficient at interaction
term CSR_REPORT 9 POST is not statistically significant,
indicating that there is no (positive) reaction/change among non-
professional stakeholders in terms of reputation around the
publication of a CSR report. CSR_PERFORMANCE is not sta-
tistically significant any more what can be caused by the smaller
variation of explanatory variables because of the reduction of the
sample size. MEDIA_EXPOSURE and R&D continue to be
positively statistically significant. Furthermore, YIELD is posi-
tive and statistically significant now.
In contrast to prior assertions, our findings indicate that
additional voluntary information in stand-alone CSR
reports does not lead to the desired result of an enhanced
reputation among non-professional stakeholders.
Additional Analyses and Robustness Checks
Additional Analyses
Investigation of Professional Stakeholders
Since stakeholder groups differ from each other concerning
their information needs and the information sources they
use (Du et al. 2010), we also consider reputational
assessments of professional stakeholders. Therefore, we
use an alternative measure for corporate reputation as
dependent variable in our Eq. (1) instead of the BrandIn-
dex.25 In line with Fombrun and Shanley (1990), we use
the Fortune magazine’s annual America’s Most Admired
Companies (AMAC) ranking after controlling for a
24
Please note that there are small variations in numbers of
observations for the treatment and control group, since some of the
matched pairs show different daily reputation data availability. This
leads to an uneven number of observations for some of the event
windows. Nevertheless, the sample sizes of the groups are comparable
and reveal a similar coverage.
25
In this analysis, we exclude MEDIA_EXPOSURE, since this
variable is available on a daily basis, whereas the alternative
reputation measure and all other variables are only available on a
yearly basis.
REPUTATION CSR_REPORT CSR_PERFORMANCE
-3.500 0.000 67.890
10,429 10,429
58.583 0.215 52.502
61.100 0.000 51.670
financial halo effect (Brown and Perry 1994).26 The scores
are based on a survey of executives, directors, and buy- and
sell-side analysts. The respondents are asked to rate leading
firms within different industry groups on the following
eight attributes: financial soundness; long-term investment
value; wise use of corporate assets; innovativeness; ability
to attract, develop, and keep talented people; quality of
products or services; quality of management; and com-
munity and environmental responsibility (Fortune 2006).
Data are available on a yearly basis resulting in 369 firm-
year observations. CSR_REPORT is an indicator variable
which is equal to 1 if a firm publishes a CSR report in year t,
and 0 otherwise. Moreover, we include an interaction of
CSR_REPORT and CSR_PERFORMANCE. Table 6, Col-
umn I (II) presents the results without (with) the interaction
term. The coefficient at CSR_REPORT is not statistically
significant, whereas the interaction term is significantly posi-
tive at p \ 10 %, indicating that professional stakeholders
evaluate the issuance of CSR reports positively in terms of
reputation if the regarded firm also has a good CSR perfor-
mance. This outcome confirms the results of prior investiga-
tions that report a positive association between environmental
disclosure and environmental reputation among experts
(Toms 2002; Cho et al. 2012; Pérez et al. 2015).
Investigation of Other CSR Disclosure Sources
Since it is unlikely that non-professional stakeholders do
not evaluate CSR information positively at all, we expect
26
Since the reputation measure could potentially suffer from a
financial halo effect that Brown and Perry (1994) documented with
regard to the AMAC ranking, we run again a regression with Brown
and Perry’s (1994) financial performance variables. We regress the
reputation measure, represented by the AMAC ranking, on market to
book ratio, average return on assets, sales, growth (percentage change
in sales), and risk (debt/equity). All financial performance variables
are highly significant and the adjusted R2 is relatively high. Therefore,
we apply the procedure presented by Brown and Perry (1994), using
the residuals of the financial halo variables regression as the
dependent variable/measure for reputation in our main regression,
hence, controlling for the financial halo effect.
123
123
Table 3 Pearson/Spearman correlations
Variable 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

1. REPUTATION 0.04 0.23 -0.29 0.31 0.36 0.02 0.09 -0.11 -0.09 0.12 0.66 0.06 0.31 0.40
2. CSR_REPORT 0.04 0.57 0.34 -0.02 0.03 0.35 -0.40 -0.07 0.24 0.25 0.06 0.17 -0.09 0.12
3. CSR_PERFORMANCE 0.19 0.55 0.46 0.11 0.08 0.48 -0.37 -0.07 0.32 0.41 0.22 0.22 -0.03 0.32
4. SIZE -0.36 0.32 0.47 -0.40 -0.39 0.33 -0.43 0.13 0.45 0.35 -0.13 0.20 -0.40 -0.01
5. ROA 0.23 -0.03 0.06 -0.29 0.60 0.09 0.01 -0.29 -0.07 0.12 0.30 -0.22 0.29 0.20
6. MtB 0.10 0.03 0.03 -0.12 0.17 0.07 0.10 -0.26 -0.16 0.06 0.30 -0.13 0.18 0.27
7. YIELD 0.05 0.32 0.38 0.27 0.11 0.02 -0.31 -0.10 0.26 0.24 0.01 0.17 -0.09 0.04
8. RISK 0.08 -0.28 -0.31 -0.24 -0.06 0.05 -0.18 0.08 -0.23 -0.12 0.09 -0.08 0.13 -0.19
9. INSTITUTIONAL -0.06 -0.09 -0.05 0.04 -0.13 -0.01 -0.15 0.06 0.04 -0.17 -0.13 0.07 0.00 -0.17
10. DIVERS -0.10 0.24 0.32 0.46 -0.01 -0.06 0.23 -0.17 0.00 0.18 -0.07 -0.07 -0.18 0.05
11. ADVERTISING 0.12 0.25 0.40 0.33 0.09 -0.06 0.22 -0.02 -0.13 0.18 0.30 0.22 0.00 0.14
12. MEDIA_EXPOSURE 0.64 0.06 0.19 -0.17 0.22 0.06 0.03 0.07 -0.07 -0.06 0.31 0.06 0.29 0.17
13. LEV 0.01 0.09 0.09 0.14 -0.19 -0.25 0.15 0.03 0.05 -0.08 0.16 0.03 -0.18 0.05
14. LITIGATION 0.34 -0.09 -0.05 -0.40 0.22 0.05 -0.10 0.08 0.13 -0.18 0.00 0.27 -0.18 -0.07
15. R&D 0.31 0.01 0.11 -0.09 0.03 0.10 -0.02 -0.09 -0.08 -0.06 0.06 0.13 -0.09 0.05
This table shows correlations for all variables. Pearson (Spearman) correlations are below (above) the diagonal. Statistical significances at least at the 0.10 level are denoted in bold print.
Significances are evaluated based on the t test (p values for the t-statistic are two-tailed). We winsorized all continuous variables at the 1 and 99 % level. All variables are defined as described in
Appendix Table 9
A. Axjonow et al.
 Table 4 Multiple regression of corporate reputation on CSR reporting
Dependent variable = REPUTATION 3 Months 2 Months 1 Month 10 Days 5 Days
I II I II I II I II I II

CSR_REPORT -0.170 3.474 -0.290 3.530 -0.286 3.462 -0.283 3.774 -0.011 4.542
(-0.11) (1.03) (-0.21) (1.12) (-0.22) (1.15) (-0.23) (1.24) (-0.01) (1.59)
CSR_PERFORMANCE 0.224 0.224 0.224 0.224 0.224
(3.12)*** (3.11)*** (3.10)*** (3.09)*** (3.08)***
HIPERFORM 5.298 5.283 5.228 5.186 5.174
(1.95)* (1.94)* (1.91)* (1.89)* (1.88)*
CSR_REPORT 9 HIPERFORM -3.098 -3.456 -3.509 -3.987 -4.677
(-0.88) (1.04) (-1.10) (-1.26) (-1.54)
SIZE -1.844 -0.873 -1.844 -0.871 -1.844 -0.871 -1.844 -0.870 -1.844 -0.870
(-1.28) (-0.56) (-1.28) (-0.56) (-1.28) (-0.56) (-1.28) (-0.56) (-1.28) (-0.56)
ROA -25.205 -24.504 -25.211 -24.639 -25.195 -24.763 -25.183 -24.843 -25.176 -24.859
(-1.26) (-1.16) (-1.26) (-1.16) (-1.26) (-1.17) (-1.26) (-1.17) (-1.26) (-1.17)
MtB -0.167 -0.153 -0.167 -0.152 -0.168 -0.152 -0.168 -0.152 -0.168 -0.152
(-1.38) (-1.27) (-1.38) (-1.27) (-1.38) (-1.26) (-1.38) (-1.27) (-1.38) (-1.27)
YIELD 3.218 4.005 3.218 4.012 3.217 4.014 3.216 4.015 3.216 4.015
(1.19) (1.39) (1.19) (1.39) (1.19) (1.39) (1.19) (1.39) (1.19) (1.39)
RISK 1.999 2.195 1.999 2.197 2.000 2.205 2.001 2.209 2.001 2.210
(0.81) (0.86) (0.81) (0.86) (0.81) (0.87) (0.81) (0.87) (0.81) (0.87)
INSTITUTIONAL 0.066 0.026 0.066 0.025 0.066 0.024 0.066 0.024 0.067 0.024
The Impact of Corporate Social Responsibility Disclosure on Corporate Reputation…

(0.97) (0.38) (0.97) (0.37) (0.97) (0.36) (0.97) (0.35) (0.97) (0.35)
DIVERS 0.443 0.474 0.443 0.478 0.443 0.481 0.443 0.483 0.443 0.484
(0.55) (0.57) (0.55) (0.58) (0.55) (0.58) (0.55) (0.58) (0.55) (0.58)
ADVERTISEMENT -5.243 -3.838 -5.244 -3.826 -5.244 -3.815 -5.244 -3.807 -5.244 -3.805
(-1.44) (-1.06) (-1.44) (-1.06) (-1.44) (-1.05) (-1.44) (-1.05) (-1.44) (-1.05)
MEDIA_EXPOSURE 1.287 1.327 1.287 1.328 1.287 1.327 1.287 1.327 1.287 1.327
(12.37)*** (12.05)*** (12.37)*** (12.04)*** (12.37)*** (12.04)*** (12.37)*** (12.04)*** (12.37)*** (12.04)***
LEV -12.094 -12.732 -12.093 -12.707 -12.096 -12.675 -12.097 -12.662 -12.098 -12.659
(-1.25) (-1.27) (-1.25) (-1.27) (-1.25) (-1.27) (-1.25) (-1.27) (-1.25) (-1.27)
LITIGATION 3.839 5.207 3.840 5.210 3.839 5.208 3.838 5.211 3.838 5.211
(0.60) (0.79) (0.60) (0.79) (0.60) (0.79) (0.60) (0.79) (0.60) (0.79)
RD 173.692 159.596 173.697 159.473 173.691 159.363 173.683 159.333 173.680 159.323
(2.24)** (1.97)* (2.24)** (1.97)* (2.24)** (1.96)* (2.24)** (1.96)* (2.24)** (1.96)*
CONSTANT 39.793 38.097 39.794 38.165 39.797 38.254 39.797 38.320 39.797 38.338
(1.73)* (1.45) (1.73)* (1.45) (1.73)* (1.45) (1.73)* (1.45) (1.73)* (1.45)

123
Industry fixed effects Included Included Included Included Included Included Included Included Included Included
 A. Axjonow et al.

include fixed effects for industry. Standard errors are clustered by firm. All variables are defined as described in Appendix Table 9. We winsorized all continuous variables at the 1 and 99 %
*, **, and *** represent significance at the 0.10, 0.05, and 0.01 level, respectively, in two-tailed t tests. Numbers reported are regression coefficients with t values (in parentheses). This

CSR_REPORT is an indicator variable which is equal to 1 if a firm publishes a CSR report on day t and on up to 90 days (3 months), 60 days (2 months), 30 days (1 month), 10 days, and 5 days
after the publication date, respectively, and 0 otherwise. The table shows the results for the different event windows (3, 2, 1 months and 10, 5 days), respectively. Reported regression models
table presents the regression results examining the effect of CSR reporting (as indicated by the coefficient at CSR_REPORT) on the measure of corporate reputation (REPUTATION).
that for these groups other CSR disclosure sources like the

114,296
firms’ websites are a more appropriate reputation

0.65
II enhancement tool compared to stand-alone CSR reports
which might be more important to financial stakeholders.
For this reason, we replace the variable CSR_REPORT in
114,296
Eq. (1) by an indicator variable that indicates whether a firm
5 Days

publishes CSR information on its website (WEBSITE).27

0.66
I

We obtain the past information of firms’ websites from the

Internet Archive Wayback Machine (http://archive.org).
This archive provides free public access to collections of
114,296

digitized materials, including websites, software applica-

0.65
II

tions/games, books, movies, music, and more. Hence, it is

possible to access the websites in the visual form they were
available during our observation period. Moreover, we
10 Days

114,296

include some additional variables into Eq. (1) to analyze

0.66

whether the optical representation of CSR information on a

I

website and the CSR information content have an impact on

reputation. These variables are OPTICS, ENVIRON-
114,296

MENTAL, and SOCIAL. OPTICS describes the optical

0.65

representation of CSR information. This variable can take

II

values between 0 and 3, whereby the value 0 is taken when

no information is published and the value 3 is taken when
1 Month

114,296

the information is presented in an optically highly appealing

0.66

way by, e.g., using a high amount of graphical representa-

I

tions. ENVIRONMENTAL and SOCIAL are indicator vari-

ables which indicate whether the website contains
information concerning environmental or social aspects.
114,296

Table 7 presents the empirical results. The final sample

0.65
II

consists of 78,592 observations, since some of our sample

firms’ websites were not available in the Internet Archive.
2 Months

To avoid the loss of a great amount of observations, we fill

114,296

missing values with values of the following year, since

0.66

there are very few changes in the construction of websites

I

during the observation period, suggesting that the structure

of websites in our sample remains constant. The coefficient
114,296

at WEBSITE is significantly positive at p \ 5 %, indicating

0.65

that CSR information provided on a firm’s homepage has a

II

positive impact on corporate reputation and, thus, can be

considered as a reputation enhancement tool for non-pro-
3 Months

114,296

fessional stakeholders.28 This implies that the CSR dis-

0.66

closure source plays a considerable role in influencing

I

stakeholders’ perceptions. OPTICS, ENVIRONMENTAL,

Dependent variable = REPUTATION

27
This binary variable indicates whether a firm publishes CSR
information on its website and this information is directly observable
for a stakeholder who visits the website. Hence, we do not classify
firms as CSR information publishing firms when information is not
directly observable for a stakeholder who visits the website, but is
somewhere ‘‘hidden’’ on the website. We define this condition, since
Table 4 continued

we assume that consumers or the general public would not spend

much time for searching CSR information on a firm’s website when
this information is not directly accessible.
28
In line with this result, Eberle et al. (2013) provide experimental
Adj. R2

level

evidence that communication about CSR through interactive online

media is beneficial to a company’s reputation.
N

123
The Impact of Corporate Social Responsibility Disclosure on Corporate Reputation…

Table 5 Multivariate differences-in-differences regression of corporate reputation on CSR reporting around the publication date of CSR reports
Dependent variable = REPUTATION 3 Months 2 Months 1 Month 10 Days 5 Days

CSR_REPORT 0.395 0.495 0.780 1.504 1.352

(0.11) (0.14) (0.22) (0.41) (0.36)
POST -0.017 -0.010 0.117 0.262 -0.063
(-0.03) (-0.02) (0.20) (0.29) (0.06)
CSR_REPORT x POST 0.138 -0.038 -0.180 -0.831 -0.205
(0.18) (-0.05) (-0.23) (-0.73) (-0.14)
CSR_PERFORMANCE 0.115 0.114 0.117 0.116 0.128
(1.19) (1.17) (1.21) (1.17) (1.31)
SIZE -1.029 -1.046 -1.062 -1.009 -1.325
(-0.58) (-0.58) (-0.60) (-0.57) (-0.77)
ROA -18.827 -19.688 -20.826 -18.742 -21.166
(-0.72) (-0.75) (-0.79) (-0.71) (-0.76)
MtB -0.200 -0.222 -0.204 -0.168 -0.214
(-0.31) (-0.35) (-0.31) (-0.26) (-0.33)
YIELD 6.137 6.063 6.118 5.961 5.796
(1.90)* (1.88)* (1.91)* (1.79)* (1.81)*
RISK 4.387 4.446 4.201 4.007 4.451
(1.26) (1.26) (1.20) (1.16) (1.26)
INSTITUTIONAL -0.150 -0.149 -0.128 -0.115 -0.089
(-1.50) (-1.50) (-1.27) (-1.16) (-0.88)
DIVERS 0.137 0.193 0.163 0.115 0.105
(0.14) (0.19) (0.17) (0.12) (0.11)
ADVERTISEMENT -5.318 -5.512 -5.507 -5.060 -4.447
(-1.17) (-1.20) (-1.22) (-1.11) (-0.98)
MEDIA_EXPOSURE 1.371 1.372 1.361 1.351 1.353
(10.53)*** (10.49)*** (10.49)*** (10.03)*** (10.10)***
LEV -18.888 -19.740 -21.193 -22.659 -20.093
(-1.38) (-1.42) (-1.56) (-1.67)* (-1.46)
LITIGATION 4.592 4.043 3.072 6.174 7.174
(0.60) (0.52) (0.40) (0.82) (0.94)
RD 194.405 188.646 193.343 214.884 229.528
(1.95)* (1.93)* (2.01)** (2.30)** (2.29)**
CONSTANT 38.522 39.680 40.049 39.057 41.942
(1.41) (1.43) (1.47) (1.43) (1.54)
Industry fixed effects Included Included Included Included Included
2
Adj. R 0.66 0.66 0.66 0.66 0.66
N 33,173 22,702 11,787 4066 2004
*, **, and *** represent significance at the 0.10, 0.05, and 0.01 level, respectively, in two-tailed t tests. Numbers reported are regression
coefficients with t values (in parentheses). This table presents the differences-in-differences regression results examining the effect of CSR
reporting (as indicated by the coefficient at CSR_REPORT x POST) around the publication date of a CSR report on the measure of corporate
reputation (REPUTATION) for different event windows (three, two, one months and 10, 5 days around the publication of a CSR report),
respectively. In this regression, CSR_REPORT is an indicator variable which is equal to 1 if a firm publishes a CSR report for the whole regarded
event window, and 0 otherwise. POST is an indicator variable which is equal to 1 if an observation relates to a post-publication date in the
regarded event window and 0 if it relates to a pre-publication date in the regarded event window for all matched reporter and non-reporter pairs
(e.g. POST equals 1 for the publication date and for 90 firm-day observations after publication, and equals 0 for 90 firm-day observations before
the report was published). Observations that exceed the regarded event window are excluded. The interaction of CSR_REPORT and POST
effectively measures the difference in changes in corporate reputation for the firms that publish a stand-alone CSR report (treatment group)
relative to non-CSR reporters (control group). Reported regression models include fixed effects for industry. We use propensity score matching
by size and industry-year to allocate publication dates of reporters to non-reporters. Standard errors are clustered by firm. All variables are
defined as described in Appendix Table 9. We winsorized all continuous variables at the 1 and 99 % level.

123

and SOCIAL have no significant impact on reputation,
indicating that information representation and content do
not play a role for non-professional stakeholders’ assess-
ments as long as generally some CSR information is
accessible on the homepage.
Environmentally Sensitive Industries
In a further test, we aim to explore whether firms which
operate in an environmentally sensitive industry experience a
reputation improvement by issuing a CSR report, to capture
the different characteristics of firms involved in environ-
mentally sensitive and non-sensitive activities (Toms 2002).
Thus, we define an indicator variable which is equal to 1 if a
firm operates in an environmentally sensitive industry group
like Mining, Oil and Gas, Paper and Pulp, Chemicals, Metals,
and Utilities (two-digit SIC codes of 10, 12, 13, 20, 29, 26, 28,
33, 49), and 0 otherwise (Patten 2002; Toms 2002; Cho et al.
2012). We include an interaction term of CSR_REPORT and
the industry indicator in our base model in Eq. (1) for all five
event windows. Again, for all event windows the interaction
term is not statistically significant and the main results remain
unchanged supporting our findings that by issuing CSR
reports firms are not able to improve their reputation among
non-professional stakeholders, even if they operate in envi-
ronmentally sensitive industries.
Attributes of the CSR Report
Moreover, we consider whether some attributes of a stand-
alone CSR report like its length and whether the report is
assured by a third party have an impact on reputation. On
the one hand, because longer reports are likely to comprise
more information indicating better transparency (Dhaliwal
et al. 2012), this could lead to better stakeholder evalua-
tions. On the other hand, the public could be overwhelmed
by too much information in longer reports not being able to
comprehend and evaluate the provided information (Par-
guel et al. 2011). This, in turn, would not lead to a repu-
tation enhancement. An assured report in turn enhances
credibility of the information provided, which might have a
positive impact on reputation (Simnett et al. 2009). Con-
cerning the length, we construct an indicator variable that
equals 1 if the number of pages of the firm’s CSR report is
greater than the median of pages of all CSR reports, and 0
otherwise (Dhaliwal et al. 2012).29 Moreover, we define an
indicator variable which equals 1 if the firm’s CSR report
29
In an additional analysis, we construct an indicator variable that
equals 1 if the number of pages of the firm’s CSR report is greater
than the 73th percentile of pages of all CSR reports, and 0 if the
number of pages is less than the 27th percentile. For constructing this
indicator variable, we follow the rule recommended by Kelly (1939)
and verified by Curento (1957), which states that upper and lower
123
A. Axjonow et al.
was assured by a third party, and 0 otherwise. We include
two additional interactions of CSR_REPORT and the two
new variables in our base model in Eq. (1). The results show
that both interactions are not statistically significant for all
five event windows. Concerning the reports’ length, our
result is consistent with the assertions by Parguel et al.
(2011). However, the finding that an assured report does not
add to reputation is tentative, given the very small number of
assured reports in our sample. Most importantly, our main
interest variable, CSR_REPORT, continues to be statistically
insignificant with the inclusion of the two new interactions.
Different Dimensions of CSR Performance
Further, we want to consider the effect of the four CSR
performance pillars (economic, governance, environmen-
tal, and social pillar) on corporate reputation separately.
For this reason, we replace CSR_PERFORMANCE by each
of the four pillars, respectively, in our base model in
Eq. (1) for all five event windows. The results reveal that
the economic, environmental, and social pillars have a
positive and significant impact on corporate reputation
(p \ 0.05, p \ 0.05, and p \ 0.1), whereas the governance
pillar has no significant effect on reputation.
Robustness Checks
Two-Stage Regressions and Self-Selection
Potentially, our study could suffer from a self-selection prob-
lem because disclosure decisions are unlikely to be random
events, as they may coincide with firms’ economic and gov-
ernance characteristics like profitability or the transparency of
their financial disclosure (Healy and Palepu 2001). Even
though we use a propensity score matching and control in the
base model for numerous factors associated with disclosure
decisions, it still could be possible that this does not control for
the self-selection problem adequately. For this reason, we also
implement the Heckman (1976) two-stage approach for pro-
viding further evidence of the robustness of our results.
For probit regression in the first-stage estimation, we
mainly use variables provided by Dhaliwal et al. (2012).
Hence, we use CSR_PERFORMANCE, SIZE, ROA, LEV, RD,
which were already defined, and market share (MKTSHR) that
is measured by a firm’s fraction of sales in its industry.
Footnote 29 continued
groups should consist of 27 % from the extremes of the criterion
score distribution. Hence, CSR reports with an amount of pages
greater than 65 were classified as high page reports and CSR reports
with an amount of pages less than 26 were classified as low page
reports. CSR reports with 26–65 pages were excluded. The results are
the same as using the rule of Dhaliwal et al. (2012).
The Impact of Corporate Social Responsibility Disclosure on Corporate Reputation…

Table 6 Analysis of corporate reputation among professional Table 7 Analysis of firm websites as an alternative CSR information
stakeholders source
Dependent variable = REPUTATION I II Dependent variable = Reputation

CSR_REPORT 0.157 -0.850 WEBSITE 22.884 (2.17)**

(1.11) (-1.53) OPTICS -5.879 (-1.30)
CSR_PERFORMANCE 0.007 0.006 ENVIRONMENTAL 7.902 (1.17)
(2.01)** (1.55) SOCIAL -17.512 (-1.42)
CSR_REPORT 9 CSR_PERFORMANCE 0.012 CSR_PERFORMANCE 0.268 (2.43)**
(1.82)* SIZE -1.790 (-0.79)
SIZE 0.011 -0.000 ROA -29.601 (-1.10)
(0.17) (-0.01) MtB -0.224 (-1.42)
ROA -0.796 -0.785 YIELD 1.653 (0.56)
(-0.75) (-0.75) RISK 0.075 (0.87)
MtB -0.001 -0.003 INSTITUTIONAL 1.102 (0.43)
(-0.10) (-0.21) DIVERS 0.652 (0.59)
YIELD 0.307 0.282 ADVERTISEMENT -5.029 (-1.03)
(3.31)*** (3.08)*** MEDIA_EXPOSURE 1.254 (9.35)***
RISK 0.002 0.001 LEV -11.681 (-0.94)
(0.59) (0.40) LITIGATION 6.389 (0.81)
INSTITUTIONAL -0.182 -0.162 RD 155.698 (1.37)
(-1.44) (-1.31) CONSTANT 39.693 (1.15)
DIVERS -0.054 -0.050 Industry fixed effects Included
(-1.27) (-1.22) Adj. R2 0.65
ADVERTISEMENT 0.048 0.041 N 78,592
(0.28) (0.24) *, **, and *** represent significance at the 0.10, 0.05, and 0.01 level,
LEV -0.835 -0.747 respectively, in two-tailed t tests. Numbers reported are regression
(-1.86)* (-1.67)* coefficients with t values (in parentheses). This table presents the
regression results examining the effect of CSR information on firms’
LITIGATION -0.467 -0.471
websites on the measure of corporate reputation (REPUTATION) rep-
(-1.01) (-1.03) resented by the YouGov BrandIndex. WEBSITE is an indicator variable
RD 0.196 0.409 that indicates whether a firm publishes CSR information on its website.
(0.05) (0.11) OPTICS describes the optical representation of CSR information. This
variable can take values between 0 and 3. The value 0 is taken when no
CONSTANT -0.466 -0.255
information is pub-lished, the value 1 is taken when information is only
(-0.45) (-0.24) provided in text form without graphical elements, the value 2 is taken
Industry fixed effects Included Included when information is accompanied by a low number of graphical repre-
Adj. R2 0.23 0.24 sentations, and the value 3 is taken when the information is presented in
an optically highly appealing way by e.g. using a high amount of
N 369 369 graphical representations. ENVIRONMENTAL and SOCIAL are indi-
*, **, and *** represent significance at the 0.10, 0.05, and 0.01 level, cator variables which indicate whether the website contains information
respectively, in two-tailed t tests. Numbers reported are regression concerning environmental or social aspects. Reported regression models
coefficients with t values (in parentheses). This table presents the include fixed effects for industry. Standard errors are clustered by firm.
regression results examining the effect of CSR reporting (as indicated All variables are defined as described in Appendix Table 9. We win-
by the coefficient at CSR_REPORT) on the measure of corporate sorized all continuous variables at the 1 and 99 % level.
reputation (REPUTATION) using the AMAC ranking. Data is
available on a yearly basis resulting in 369 firm-year observations.
regression. In the second-stage, we add the inverse mills ratio.
CSR_REPORT is an indicator variable which is equal to 1 if a firm
publishes a CSR report in year t, and 0 otherwise. Column I (II) Our main results are unchanged for all five event windows.
presents the results without (with) the interaction term of CSR_RE-
PORT and CSR_PERFORMANCE. Reported regression models Alternative Dependent Variable and Control Variables
include fixed effects for industry. Standard errors are clustered by
firm. All variables are defined as described in Appendix Table 9. We
winsorized all continuous variables at the 1 and 99 percent level. We conduct some robustness tests using alternative defi-
nitions for our original variables. Our major results remain
Moreover, we include MtB and LITIGATION (Dhaliwal et al. the same using sales as proxy for size, using return on
2011). The results show that most variables are statistically equity instead of return on assets, and using the ratio of
significant and have the predicted sign in the probit-model total debt to common equity as proxy for leverage.

123

Furthermore, we standardize REPUTATION. With this
variation, we obtain the same results as those reported in
Table 4: there is no statistically significant association
between CSR reporting and corporate reputation and, thus,
no resulting reputation enhancement.
Moreover, we use an alternative method to calculate our
original measure (REPUTATION) by including neutral
respondents in the calculation. This means that we obtain
the daily scores by dividing the difference of positive and
negative respondents by the number of all respondents
(positive ? negative ? neutral respondents). This alterna-
tive proxy and our original proxy are highly correlated with
one another, which suggest that they probably cover the
same underlying construct of reputation. All else equal,
using this new proxy, we obtain results qualitatively similar
to those based on REPUTATION. We also construct an
average measure of the two reputation proxies because this
potentially reduces noise in individual measures (Larcker
and Rusticus 2010). Here again, the issuance of CSR
reports has no impact on corporate reputation.
Discussion
This paper investigates the impact of CSR disclosure on cor-
porate reputation as perceived by non-professional stakehold-
ers. The results of our empirical analyses show that stand-alone
CSR reporting has no impact on a firm’s reputation when
considering non-professional stakeholders like (actual or
potential) customers, employees, retail investors, and the gen-
eral public. This result holds for the long-term analyses as well
as for the short-term analyses, when examining corporate rep-
utation around the dates of CSR report publication. This finding
supports our arguments that CSR reporting may be seen as a
waste of resources (Izzo and Magnanelli 2012), there may exist
a lack of credibility concerning the content of CSR reports
(Simnett et al. 2009; Moser and Martin 2012), and the reports
may be apprehended by stakeholders as an ingratiating attempt
to gain public favor (Godfrey 2005), thus, having no impact on a
firm’s reputation.
In contrast to prior assertions, our finding indicates that
additional voluntary information concerning CSR activities
and, thus, the firms’ attempt to provide additional information
by issuing CSR reports in terms of more transparency, does not
lead to the desired result of an enhanced reputation, respecting
perceptions of non-professional stakeholders. Regarding CSR
performance, we provide evidence that corporate reputation is
significantly positively influenced by firms’ CSR performance.
This result is in line with prior literature (Fombrun and Shanley
1990; Williams and Barrett 2000; Sen and Bhattacharya 2001;
Luo and Bhattacharya 2006; Lev et al. 2010).
However, the absence of a statistically significant
association between financial indicators like financial
123
A. Axjonow et al.
performance or risk and corporate reputation seems to be
surprising in view of Fombrun and Shanley (1990), who
report significant results. This could be due to the fact that
the respondents in our setting represent the general popu-
lation and, thus, non-professional stakeholder groups like
consumers and the general public. Using the AMAC
rankings as a measure for reputation, Fombrun and Shanley
(1990) rely only on evaluations of experts like executives,
directors, and buy- and sell-side analysts, who are very
well informed about the financial situation of the compa-
nies they are asked to evaluate. Furthermore, the firms’
financial success is more relevant for these expert groups
than for consumers or the general public, which can lead to
the fact that their evaluations are more likely to be driven
by financial aspects than the evaluations of non-expert
stakeholders.
The circumstance that contrary to the AMAC rankings or
Newsweek magazine rankings, used by Cho et al. (2012),
non-experts evaluate the firms, could also be important in
order to explain our main result of CSR reporting having no
impact on reputation. Experts may be more informed about a
company’s reporting activities and also more interested in
additional information provided explicitly by CSR reports
than consumers or the general public. Non-professional
stakeholders might simply not know that a firm publishes a
CSR report or not be able to deal with the overwhelming
amount of information provided in such reports (Parguel
et al. 2011). Hence, CSR information provided by another
medium could be more appropriate when addressing non-
professional stakeholders. These considerations show that it
is very important to differentiate between professional and
non-professional stakeholders and the CSR information
sources they use. Du et al. (2010) also argue that stakeholder
groups differ from each other concerning their information
needs and the information channels they use. Thus, we also
consider reputational assessments of professional stake-
holders and the firm’s websites as an alternative CSR
information channel. In line with prior research, we find that
stand-alone CSR reports have a positive impact on reputa-
tion among professional stakeholders if the firms also have a
good CSR performance. Regarding the firm’s websites, we
show that non-professional stakeholders react positively in
terms of reputation to CSR information published through
this channel. In summary, we provide evidence that stake-
holder group and CSR disclosure channel play an important
role in influencing reputational perceptions.
Conclusions, Limitations, and Future Research
We examine one of the most frequently mentioned benefits
by CSR disclosure proponents: namely that CSR disclosure
leads to an enhancement of corporate reputation. Contrary to
The Impact of Corporate Social Responsibility Disclosure on Corporate Reputation…

this view, we find that the issuance of stand-alone CSR
reports is not significantly associated with firms’ reputation
among non-professional stakeholders. Further analyses
confirm our findings. Additional analyses reveal that the
publication of stand-alone CSR reports leads to a reputation
improvement among professional stakeholders if the firms
also have a good CSR performance. Moreover, non-profes-
sional stakeholders react positively in terms of reputation to
CSR information published on the firm’s website. These
results imply that stand-alone CSR reports can be used as a
tool for reputation improvement among professional stake-
holders, but non-professional stakeholders are reached by
using the firm’s website as a CSR disclosure source.
This study contributes to the CSR disclosure literature
by extending the research on CSR disclosure beyond the
focus on financial stakeholders’ reactions, analyzing the
perceptions of non-professional stakeholders in terms of
corporate reputation. Moreover, we shed light on the
entrenched view that stand-alone CSR reporting should
generally enhance a firm’s reputation. Our investigation
provides a unique empirical analysis of non-professional
stakeholder assessments, since the YouGov BrandIndex
allows us to examine daily firm perceptions of non-ex-
perts like consumers, employees, and the general public,
making it possible to conduct an event-based analysis
with reputational measures. This can be regarded as a
novelty, as generally daily event analyses are based on
capital market information. Hence, our study provides an
important empirical contribution to the theoretical dis-
cussion of the presented question. Moreover, this study
offers practical implications for companies in making
decisions for or against stand-alone CSR reporting and
about the location of CSR information concerning dif-
ferent stakeholder groups. Firms can draw on the pre-
sented results when making decisions concerning the
disclosure channel to be favored dependent on the
stakeholder group that is addressed. Moreover, firms can
benefit from the presented findings, as they can invest in
CSR disclosure in a more targeted way to achieve repu-
tational improvements.
We emphasize that our investigation is not free of lim-
itations. The use of the differences-in-differences approach
in our study is not based on an exogenous shock, since
firms can decide on their own, whether to publish a vol-
untary CSR report. However, we control for a possible self-
selection bias by using a propensity score matching and by
the Heckman two-stage approach. Furthermore, by con-
centrating on the issuance of CSR reports only, we fail to
capture the effects of the content of reports on reputation.
Further, corporate reputation is a construct that has to be
measured by using proxies. Although we believe that our
reputation measure is a valid proxy, it might not fully
reflect the underlying concept of reputation. Regarding our
YouGov data, a possible limitation is that respondents
evaluate the brands with only positive, negative, and neu-
tral options, thus, allowing no ratings on an interval scale
(Luo et al. 2013). Finally, our sample is limited to a con-
sideration of stakeholder perceptions only in the U.S.
market for a short time period, which may raise concerns
about the validity of our results.
Overall, we believe that the investigation of how groups
of non-professional stakeholders perceive CSR disclosure
in terms of reputation is an important enrichment of the
existing CSR literature and we hope that this will spur
future research into how also other stakeholders than
investors react to CSR disclosures. Moreover, future
research can examine how further CSR information sour-
ces, beyond CSR reports and firm websites, influence
corporate reputation. Finally, it would also be of particular
interest to investigate the effect of the information content
or quality of stand-alone CSR reports on non-professional
stakeholders’ perceptions and, thus, on reputation.
Compliance with Ethical Standards
Conflict of interest The authors declare that they have no conflict of
interest.
Ethical approval This article does not contain any studies with
human participants or animals performed by any of the authors.
Appendix
See Tables 8, 9 and Fig. 1.

Table 8 YouGov BrandIndex dimensions

Dimension Question

Perceived brand quality
Perceived brand value
Perceived brand satisfaction
Perceived brand recommendation
Perceived brand affect
Perceived brand-workplace reputation
‘‘Which of the brands in the sector do you associate with good or poor quality?’’
‘‘Which of the brands do you associate with good or poor value-for-money?’’
‘‘Would you identify yourself as a recent satisfied or an unsatisfied customer of any of these brands?’’
‘‘Which brands would you recommend to a friend? or suggest avoiding?’’
‘‘Of which brands do you have a ‘generally positive’ or ‘generally negative’ impression?’’
‘‘Which of the brands would you be proud/embarrassed to work for?’’

123
 A. Axjonow et al.

Table 9 Variable definitions

Variables Description

Dependent variable/proxy for corporate reputation

REPUTATION Corporate reputation score in terms of the daily aggregate BrandIndex values. Daily values are constructed by taking
the difference between positive and negative respondents divided by the sum of positive and negative respondents
CSR variables
CSR_REPORT An indicator variable which is equal to 1 if a firm publishes a CSR report, 0 otherwise
CSR_PERFORMANCE An average score of the four (environmental, social, governmental, economic) pillars provided by the Thomson
Reuters ASSET 4 dataset
HIPERFORM An indicator variable which is equal to 1 if the firm’s CSR performance is greater than the median CSR performance
of the firm’s industry peers, 0 otherwise
POST An indicator variable which is equal to 1 if an observation relates to a post-publication day in the regarded event
window of 90 days (3 months), 60 days (2 months), 30 days (1 month), 10 days, and 5 days before/after the
publication date of a CSR report, respectively, 0 otherwise
CSR_REPORT 9 POST The interaction of CSR_REPORT and POST
Control variables
SIZE The natural logarithm of the firm’s total assets at the end of the previous year, in thousand $
ROA Return on assets as the net income before interest and taxation divided by the total assets at the end of the previous
year
MtB The ratio of the market value of equity to book value of equity at the end of the previous year
YIELD Dividend yield as the ratio of dividend to share price at the end of the previous year
RISK The firm’s market risk measured by the firm’s beta coefficient as a measure of the systematic risk at the end of the
previous year
INSTITUTIONAL The firm’s stock concentration in institutional hands at the end of the previous year, in percent
DIVERS The firm’s number of segments in which it operates at the end of the previous year
ADVERTISING Advertising intensity measured by an indicator variable which equals 1 if the regarded firm was on the ‘‘100 Leading
U.S. Advertisers’’ list in 2011 regarding its advertising expenditures, monitored by Advertising Age, 0 otherwise
MEDIA_EXPOSURE Media exposure provided by YouGov in terms of daily assessments regarding negative or positive associations about
the brand in the media. Daily values are constructed by taking the difference between positive and negative
respondents divided by the sum of positive, negative, and neutral respondents
LEV Leverage as the ratio of total debt to total assets at the end of the previous year
LITIGATION An indicator variable which is equal to 1 if the firm operates in a high-litigation industry (SIC codes of 2833–2836,
3570–3577, 3600–3674, 5200–5961, and 7370) in the previous year, 0 otherwise
R&D Research and development intensity measured by the ratio of R&D expenditure to total assets at the end of the
previous year

References

Akerlof, G. (1970). The market for ‘lemons’: Quality uncertainty and

the market mechanism. Quarterly Journal of Economics, 84(3),
488–500.
Barone, E., Cunningham, G. M., & Marnet, O. (2014). Company
reputation and disclosures: Review, synthesis and analysis. In G.
M. Cunningham (Ed.), Reclaiming accounting’s lost identity (pp.
85–110). Mjölby: Atremi.
Basdeo, D. K., Smith, K. G., Grimm, C. M., Rindova, V. P., &
Derfus, P. J. (2006). The impact of market actions on firm
reputation. Strategic Management Journal, 27(12), 1205–1219.
Bebbington, J., Larrinaga, C., & Moneva, J. M. (2008). Corporate
social reporting and reputation risk management. Accounting
Auditing and Accountability Journal, 21(3), 337–361.
Berens, G., van Riel, C. B. M., & van Bruggen, G. H. (2005).
Fig. 1 Reputation distribution over time Corporate associations and consumer product responses: The

123
The Impact of Corporate Social Responsibility Disclosure on Corporate Reputation…
moderating role of corporate brand dominance. Journal of
Marketing, 69(3), 18–35.
Brammer, S. J., & Pavelin, S. (2006). Corporate reputation and social
performance: The importance of fit. Journal of Management
Studies, 43(3), 435–455.
Brown, B., & Perry, S. (1994). Removing the financial performance
halo from Fortune’s ‘‘Most Admired’’ companies. Academy of
Management Journal, 37(5), 1347–1359.
Cheng, M. M., Green, W. J., & Ko, J. C. (2015). The impact of
Strategic relevance and assurance of sustainability indicators on
investor’s decisions. Auditing, 34(1), 131–162.
Cheng, B., Ioannou, I., & Serafeim, G. (2014). Corporate social
responsibility and access to finance. Strategic Management
Journal, 35(1), 1–23.
Cho, C. H., Guidry, R. P., Hageman, A. M., & Patten, D. M. (2012).
Do actions speak louder than words? An empirical investigation
of corporate environmental reputation. Accounting Organiza-
tions and Society, 37(1), 14–25.
Cho, C. H., & Patten, D. M. (2007). The role of environmental
disclosures as tools of legitimacy: A research note. Accounting
Organizations and Society, 32(7/8), 639–647.
Clarkson, P. M., Fang, X., Li, Y., & Richardson, G. D. (2013). The
relevance of environmental disclosures: Are such disclosures
incrementally informative? Journal of Accounting and Public
Policy, 32(5), 410–431.
Clarkson, P. M., Li, Y., Richardson, G. D., & Vasvari, F. P. (2008).
Revisiting the relation between environmental performance and
environmental disclosure: An empirical analysis. Accounting
Organizations and Society, 33(4/5), 303–327.
Curento, E. (1957). The upper and lower twenty-seven per cent rule.
Psychometrika, 22(3), 293–296.
de Villiers, C., & van Staden, C. J. (2011). Where firms choose to
disclose voluntary environmental information. Journal of
Accounting and Public Policy, 30(6), 504–525.
Dhaliwal, D. S., Li, O. Z., Tsang, A., & Yang, Y. G. (2011).
Voluntary non-financial disclosure and the cost of equity capital:
The initiation of corporate social responsibility reporting. The
Accounting Review, 86(1), 59–100.
Dhaliwal, D. S., Li, O. Z., Tsang, A., & Yang, Y. G. (2014). Corporate
social responsibility disclosure and the cost of equity capital: The
roles of stakeholder orientation and financial transparency.
Journal of Accounting and Public Policy, 33(4), 328–355.
Dhaliwal, D. S., Radhakrishnan, S., Tsang, A., & Yang, Y. G. (2012).
Non-financial disclosure and analyst forecast accuracy: Interna-
tional evidence on corporate social responsibility disclosure. The
Accounting Review, 87(3), 723–759.
Du, S., Bhattacharya, C. B., & Sen, S. (2010). Maximizing business
returns to corporate social responsibility (CSR): The role of CSR
communication. International Journal of Management Reviews,
12(1), 8–19.
Dye, R. A. (1985). Disclosure of non-proprietary information. Journal
of Accounting Research, 23(1), 123–145.
Eberle, D., Berens, G., & Li, T. (2013). The impact of interactive
corporate social responsibility communication on corporate
reputation. Journal of Business Ethics, 118(4), 731–746.
Eccles, R. G., Ioannou, I., & Serafeim, G. (2014). The impact of
corporate sustainability on organizational processes and perfor-
mance. Management Science, 60(11), 2835–2857.
Fama, E., & French, K. (2015). Industry classification. Retrieved
December 09, 2015, from http://mba.tuck.dartmouth.edu/pages/
faculty/ken.french/Data_Library/det_12_ind_port.html.
Fombrun, C. J. (2001). Corporate reputation—Its measurement and
management. Thexis, 4, 23–26.
Fombrun, C. J., & Shanley, M. (1990). What’s in a name? Reputation
building and corporate strategy. Academy of Management
Journal, 33(2), 233–258.
Fombrun, C. J., & van Riel, C. (1997). The reputational landscape.
Corporate Reputation Review, 1(1/2), 5–13.
Fortune. (2006). How we conduct the most admired survey. Fortune,
153(4), 86.
Francis, J., Philbrick, D., & Schipper, K. (1994). Shareholder
litigation and corporate disclosures. Journal of Accounting
Research, 32(2), 137–164.
Global Reporting Initiative (GRI). (2016). Retrieved February 25,
2016, from https://www.globalreporting.org/information/sustain
ability-reporting/Pages/reporting-benefits.aspx.
Godfrey, P. C. (2005). The relationship between corporate philan-
thropy and shareholder wealth: A risk management perspective.
Academy of Management Review, 30(4), 777–798.
Gow, I. D., Ormazabal, G., & Taylor, D. J. (2010). Correcting for
cross-sectional and time-series dependence in accounting
research. The Accounting Review, 85(2), 483–512.
Griffin, P. A., & Sun, Y. (2013). Going green: Market reaction to
CSRwire news releases. Journal of Accounting and Public
Policy, 32(2), 93–113.
Hasseldine, J., Salama, A. I., & Toms, J. S. (2005). Quantity versus
quality: The impact of environmental disclosures on the
reputations of UK plcs. British Accounting Review, 37(2),
231–248.
Healy, P. M., & Palepu, K. G. (2001). Information asymmetry,
corporate disclosure, and the capital markets: A review of the
empirical disclosure literature. Journal of Accounting and
Economics, 31, 405–440.
Heckman, J. J. (1976). The common structure of statistical models of
truncation, sample selection and limited dependent variables and
a simple estimator for such models. Annals of Economic and
Social Measurement, 5(4), 475–492.
Hooghiemstra, R. (2000). Corporate communication and impression
management—New perspectives why companies engage in corpo-
rate social reporting. Journal of Business Ethics, 27(1), 55–68.
Huang, X., & Watson, L. (2015). Corporate social responsibility
research in accounting. Journal of Accounting Literature, 34(1),
1–16.
Hur, W.-M., Kim, H., & Woo, J. (2014). How CSR leads to corporate
brand equity: Mediating mechanisms of corporate brand cred-
ibility and reputation. Journal of Business Ethics, 125(1), 75–86.
Izzo, M. F., & Magnanelli, S. B. (2012). Does it pay or does firm pay?
The relation between CSR performance and the cost of debt.
Working paper, Luiss University.
Kelly, T. R. (1939). The selection of upper and lower groups for the
validation of test items. The Journal of Educational Psychology,
30(1), 17–24.
KPMG International. (2011). KPMG International survey of corpo-
rate responsibility reporting 2011. Amstelveen: KPMG
International.
KPMG International. (2013). KPMG International survey of corporate
responsibility reporting 2013. Amstelveen: KPMG International.
Larcker, D. F., & Rusticus, T. O. (2010). On the use of instrumental
variables in accounting research. Journal of Accounting and
Economics, 49(3), 186–205.
Larkin, Y. (2013). Brand perception, cash flow stability, and financial
policy. Journal of Financial Economics, 110(1), 232–253.
Lev, B., Petrovits, C., & Radhakrishnan, S. (2010). Is doing good
good for you? Yes, charitable contributions enhance revenue
growth. Strategic Management Journal, 31(2), 182–200.
Linthicum, C., Reitenga, A., & Sanchez, J. (2010). Social responsi-
bility and corporate reputation: The case of the Arthur Andersen
Enron audit failure. Journal of Accounting and Public Policy,
29(2), 160–176.
Luo, X., & Bhattacharya, C. B. (2006). Corporate social responsibil-
ity, customer satisfaction, and market value. Journal of Market-
ing, 70(4), 1–18.
123

Luo, X., Raithel, S., & Wiles, M. A. (2013). The impact of brand
rating dispersion on firm value. Journal of Marketing Research,
50(3), 339–415.
Luo, X., Zhang, J., Gu, B., & Phang, C. W. (2014). Competitive and
asymmetric nature of relationships between expert blog senti-
ment and general consumer brand perception. In PACIS (p. 76).
Lys, T., Naughton, J. P., & Wang, C. (2015). Signaling through
corporate accountability reporting. Journal of Accounting and
Economics, 60(1), 56–72.
Martin, P., & Moser, D. (2016). Managers’ green investment
disclosures and investors’ reaction. Journal of Accounting and
Economics, 61(1), 239–254.
Matsumura, E. M., Prakash, R., & Vera-Muñoz, S. C. (2014). Firm-
value effects of carbon emissions and carbon disclosures. The
Accounting Review, 89(2), 695–724.
Moser, D. V., & Martin, P. R. (2012). A broader perspective on
corporate social responsibility research in accounting. The
Accounting Review, 87(3), 797–806.
Nikolaeva, R., & Bicho, M. (2011). The role of institutional and
reputational factors in the voluntary adoption of corporate social
responsibility reporting standards. Journal of the Academy of
Marketing Science, 39(1), 136–157.
Panico, M., Raithel, S., & Michel, E. (2014). The effect of media
coverage on employer reputation. Journal of Media Economics,
27(4), 181–198.
Parguel, B., Benoı̂t-Moreau, F., & Larceneux, F. (2011). How
sustainability ratings might deter ‘greenwashing’: A closer look
at ethical corporate communication. Journal of Business Ethics,
102(1), 15–28.
Patten, D. M. (2002). The relation between environmental perfor-
mance and environmental disclosure: A research note. Account-
ing Organizations and Society, 27(8), 763–773.
Pérez, A., del Mar Garcia de los Salmones, M., & López, C. (2015).
Corporate reputation in the Spanish context: An interaction
between reporting to stakeholders and industry. Journal of
Business Ethics, 129(3), 733–746.
Plumlee, M., Brown, D., Hayes, R., & Marshall, S. (2015). Voluntary
environmental disclosure quality and firm value: Further
evidence. Journal of Accounting and Public Policy, 34(4),
336–361.
123
A. Axjonow et al.
Rindova, V. P., Williamson, I. O., Petkova, A. P., & Sever, J. M.
(2005). Being good or being known: An empirical examination
of the dimensions, antecedents, and consequences of organiza-
tional reputation. Academy of Management Journal, 48(6),
1033–1049.
Roberts, P. W., & Dowling, G. R. (2002). Corporate reputation
sustained superior financial performance. Strategic Management
Journal, 23(12), 1077–1093.
Sarstedt, M., Wilczynski, P., & Melewar, T. C. (2013). Measuring
reputation in global markets—A comparison of reputation
measures’ convergent and criterion validities. Journal of World
Business, 48(3), 329–339.
Schlenker, B. R. (1980). Impression management: The selfconcept, social
identity, and interpersonal relations. Belmont, CA: Wadsworth.
Schwaiger, M. (2004). Components and parameters of corporate
reputation—An empirical study. Schmalenbach Business
Review, 56(1), 46–71.
Scott, S. G., & Lane, V. R. (2000). A stakeholder approach to
organizational identity. Academy of Management Review, 25(1),
43–62.
Sen, S., & Bhattacharya, C. B. (2001). Does doing good always lead
to doing better? Consumer reactions to corporate social respon-
sibility. Journal of Marketing Research, 38(2), 225–243.
Simnett, R., Vanstraelen, A., & Chua, W. F. (2009). Assurance on
sustainability reports: An international comparison. The
Accounting Review, 84(3), 937–967.
Sweeney, L., & Coughlan, J. (2008). Do different industries report
corporate social responsibility differently? An investigation
though the lens of stakeholder theory. Journal of Marketing
Communications, 14(2), 113–124.
Toms, J. S. (2002). Firm resources, quality signals and the determi-
nants of corporate environmental reputation: Some UK evidence.
British Accounting Review, 34(3), 257–282.
Verrecchia, R. (1983). Discretionary disclosure. Journal of Account-
ing and Economics, 5(1), 179–194.
Watts, R., & Zimmerman, J. (1986). Positive accounting theory.
Englewood Cliffs, NJ: Prentice Hall.
Williams, R. J., & Barrett, J. D. (2000). Corporate philanthropy,
criminal activity and firm reputation: Is there a link? Journal of
Business Ethics, 26(4), 341–350.